downstream consequences of childhood lead poisoning€¦ · lead poisoning in young children has...

Center on Urban Poverty and Community Development

Downstream Consequences of Childhood

Lead Poisoning:

A Longitudinal Study of Cleveland Children from

Birth to Early Adulthood

Summary: Lead poisoning typically occurs in early childhood, but it can have long-term effects on ex-

posed individuals. It is important to document some of the downstream consequences of lead poison-

ing in order to appreciate the costs of inaction and to target prevention resources where most needed. In this brief, we report on a longitudinal study of two cohorts of Cleveland youth that had

elevated blood lead levels (≥ 5 µg/dl) before age 3 and matched comparison groups whose lead tests were not as elevated. We track markers of their educational success throughout elementary and high

school, and also look at adverse events, such as juvenile delinquency, adult incarceration,

homelessness, and having to rely on public assistance. We find that children with elevated lead levels

in early childhood have significantly worse outcomes on markers of school success, and higher rates

of adverse events in adolescence and early adulthood, compared to their non-exposed peers. The size

of these disparities is generally in the 20-30% range across both cohorts, and represents a sizable

societal cost due to the loss of human capital, the burden on local systems, and persistence of

inequality.

Claudia Coulton, Francisca Garcí a-Cobia n Richter, Youngmin Cho, Jiho Park, & Robert Fischer

* Updated on July 2, 2020Acknowledgment: This research was made possible through funding from Mt. Sinai Healthcare Foundation,

Saint Luke's Foundation, George Gund Foundation, and the Eva L. and Joseph M. Bruening Foundation.

Note: Lead data used in this report come from the Ohio Department of Health. This should not be considered an endorsement of this study or these conclusions by the Ohio Department of Health.

June 2020*


2

INTRODUCTION

Acting on lead poisoning prevention promises to have benefits that well exceed the costs. Much of this

return-on-investment is anticipated to come from avoidance of negative consequences of lead poisoning over

the course of child and adolescent development. Numerous studies show that there are long term costs of lead

poisoning to society in the form of lower lifetime earnings, neuropsychiatric disorders, special educational

needs, lower tax contributions, and criminal involvement.1,2 However, there is little research that traces the im-

pact of childhood lead exposure on public systems in regions where lead poisoning rates have been persistently

high and many children have been affected. In this report, we examine some of the downstream consequences

of lead poisoning in Cleveland from the perspective of the systems that serve children and youth. The focus is on

estimating the net increase in the risk of selected, and potentially costly, events in these systems that can be

attributed to lead poisoning.

BACKGROUND

Lead poisoning in young children has been recognized as a public health problem for many years. A

number of studies have shown that lead exposure in early stages of life has adverse effects on child develop-

ment, such as cognitive impairment manifested by scores on intelligence tests,3–5 poor academic achievement,6,7

and behavioral problems.8,9 A meta-analysis of 24 quantitative studies published in 1990 documented that early

childhood lead exposure impairs children’s IQ even at levels that had previously been considered safe.10 A re-

cent study in Cleveland also demonstrated that young children with blood lead levels at or exceeding the threshold of 5 µg/dl displayed lower scores on kindergarten readiness compared to their unexposed peers.11,12

Further-more, studies suggest that early childhood lead exposure may have persistent effects over a longer

period, as seen in diminished cognitive abilities, delinquency, or criminal behaviors in later adolescence and

even adult-hood.13,14

However, few studies have been able to fully evaluate the long-term consequences of early childhood

lead poisoning, due to the lack of available datasets to track its impacts on later outcomes, and the difficulties of

estimating the causal impact of lead in the presence of other factors that increase both the chances of lead ex-

posure and of poor outcomes.15 Only a few studies that attempt to examine downstream consequences of early

childhood lead poisoning have used research designs that enable causal inference. One such study in Rhode Is-

land used linked administrative data to estimate the effects of lead on school suspensions and juvenile delin-

quency for children born between 1990 and 2004.16 To isolate the causal effects of lead, this research used sib-

ling fixed effects models and instrumental variable (IV) models that exploit lead variation due to traffic exposure.

The Rhode Island study found that increased lead exposure resulted in increased school suspensions and juve-

nile detention for adolescent boys, but not for girls. Another study in Chicago, also sought to evaluate the causal

effects of early childhood lead exposure on youth development. To reduce confounding bias, these researchers

used coarsened exact matching (CEM) of children with and without elevated blood lead levels, and instrumental

variables models that leveraged lead variation from distance to a smelting plant for children born in in the

1990’s. They found a causal effect of early childhood lead exposure and later records of adolescent delinquency,

but not official arrests.17

Building on these examples, the current study examines the impact of early childhood lead poisoning

throughout childhood and the transition to adulthood, using administrative records for two cohorts of children

attending school in Cleveland, Ohio. It estimates the effect of lead poisoning by comparing outcomes among

children that had lead levels at or above the public health threshold of 5 µg/dl and a matched comparison


3

group of children whose lead values did not exceed the threshold. The differences in the two groups illustrate

the impact that lead poisoning has had on local institutions and in prolonging racial inequities. The findings are

suggestive of the societal benefits and reduction in local burden that can be anticipated as lead safety in housing

is achieved.

STUDY METHODS

We conduct a longitudinal study of two cohorts of Cleveland youth that had elevated blood lead levels

(≥ 5 µg/dl) before age 3 and matched comparison groups whose lead tests did not exceed this public health threshold. We track markers of their educational success throughout elementary and high school, and also look

at adverse events such as juvenile delinquency, adult incarceration, unemployment, homelessness and having to

rely on public assistance in early adulthood. The study sample (N=10,470) includes students that attended 9th

grade in the Cleveland metropolitan School District (CMSD) in 2007-2008 (Early cohort) or 2016-2017 (Recent

cohort). For each student, we build a longitudinal record from birth through early adulthood, drawing on admin-

istrative records from multiple systems. We use coarsened exact matching to compare the outcomes of children with and without elevated lead levels who are similar on numerous background and early childhood characteris-

tics. This process allows us to statistically estimate the impact of lead on subsequent outcomes adjusting for po-

tential confounders. The advantage of having two cohorts is that we can the estimate effects of lead exposure

for a contemporaneous group of children and young adults, and evaluate how these effects have persisted over

time.

DATA AND MEASURES

Figure 1 illustrates the general approach to building the longitudinal data for the study cohorts. Records

from all of these systems are linked together for each child in the study, using tools within the Child-Household

Integrated Longitudinal Data system maintained by the Center (See Appendix A for a detailed description of

Figure 1. Diagram of lead exposure outcomes by system and chronology

https://case.edu/socialwork/povertycenter/data-systems/child-data-system

https://case.edu/socialwork/povertycenter/data-systems/child-data-system


4

study variables and data sources). The main predictor in the study is whether the child was found to have an

elevated blood lead level in early childhood. Lead exposure is a dichotomous variable, with elevated blood lead

level (EBLL) set at the public health threshold of equal or greater than 5 µg/dl. We determine the EBLL status for each child by taking the geometric mean of all venous and capillary tests that occurred before age five. This

summary measure is less influenced by outliers than the arithmetic mean, and has been used in similar studies

looking at the developmental consequences of lead poisoning.

The outcomes for the study reflect the child’s experiences over time in the educational, judicial, employ-

ment and social services systems. The educational outcomes for the study include measures of kindergarten

readiness, proficiency test and graduation test passage, grade repetition in selected grades, and high school

graduation. All of these measures are dichotomous, and are determined based on the last score if the tests are

taken multiple times. Justice system involvement is measured based on whether or not the child was the subject

of a delinquency filing in Juvenile Court, and whether they had a record of incarceration in county jail between

the age of 18 and 23. Whether the individual is employed at age 23 is ascertained through the Ohio Wage Rec-

ord System that has records of covered quarterly employment (not including self-employed, independent con-

tractors and federal government employees). In the area of social services, we determine whether the individual

has a record of utilization of homelessness services or public assistance programs between 18 and 23 years old.

It is important to note that relying on state and local administrative data to track outcomes imposes some limi-

tations. In particular, we assume that individuals who do not have a record of an event (e.g. high school gradua-

tion, employment, homelessness) did not experience these outcomes. Although it is possible that some individu-

als are missing because they experienced the event outside of Ohio, we assume that these measurement errors

are ‘randomly’ distributed across our EBLL and non-EBLL

groups, not affecting the validity of our estimates of lead

poisoning impact.

STUDY POPULATION

Due to data availability, we restrict the study popula-

tion to children born in Ohio that had at least one lead test

by their 5th birthday. These restrictions result in a sample

comprising approximately 62% of the students that were in

the two 9th grade cohorts. Students in our early cohort were

born between 1991 and 1993, and those in the more recent

cohort were born between the years 1999 and 2002.

As can be seen in the top portion of Table 1 (Section

A), lead poisoning rates were considerably higher in the ear-

lier cohort. The geometric mean of test results averaged

across children in the early cohort was 10.6 µg/dl, almost

twice that for the recent cohort. Further, 85 percent of the children who were tested in the early cohort met the public health threshold of ≥ 5 µg/dl, compared to 48% of children in the recent cohort. With respect to demographic

characteristics, there were no differences in the two cohorts

on sex or on their mean age in 9th grade. However, we see

differences in the racial and ethnic makeup of the students

between cohorts, as the share of African American students Table 1: Lead poisoning rates and selected char-acteristics of 9th grade student cohorts

Early Cohort Late Cohort

2007-2008 2016-2017

(n = 6,063) (n = 4,414)

A: All Students

85% 48%

10.6 5.9

15.6 15.6

Lead Levels

EBLL (BLL ≥ 5µg/dl) BLL (Geometric mean)

Age (mean at 9th grade)

Sex

Female 50% 50%

Race/Ethnicity

Black/African American 82% 73%

Hispanic/Latino 6% 10%

White 10% 14%

Other 2% 2%

B: Percent students with EBLL by sex and race

Sex

Male 85% 51%

Female 84% 45%

Race/Ethnicity

Black/African American 87% 51%

Hispanic/Latino 72% 35%

White 80% 47%

Other 78% 42%


5

decreased and the shares of white and Hispanic/Latino students increased over time. In section B of Table 1,

rates of EBLL are displayed by sex and race/ethnicity. Notably, black or African American children are the most

impacted, as the rate of EBLL is highest for this group in both cohorts. Boys have higher rates of EBLL than girls

only in the recent cohort.

MATCHED COMPARISON GROUP ANALYSIS

From the above population, we create two matched groups for comparison. Following previous re-

search, we dichotomize lead levels at the public health standard of ≥ 5 µg/dl of blood to define the groups. We apply coarsened exact matching (CEM) to achieve comparability between the groups. CEM is a non-parametric

technique that temporarily coarsens values on the matching variables, finds exact matches for each case using

the coarsened data, and estimates the model on the matched, un-coarsened data.18 We match our EBLL sample children to others that do not exceed the threshold for EBLL on numerous variables (i.e., potential confounders)

that could simultaneously be related to the chances of lead exposure and to the outcomes of interest. The

matching variables include year of birth, race/ethnicity, and dichotomous variables to account for low-birth

weight status, having a teenage mother, mother without a high school diploma, mother who smoked during

pregnancy, mother who did not have pre-natal care, and mother who spoke a native language other than Eng-

lish. Additional matching variables that are dichotomized by the CEM algorithm are months the child lived in

public housing between 12 and 18 months of age, months the household received Supplemental Nutrition Assis-

tance Program (SNAP) benefits between 12 and 18 months of age, and the Opportunity Index19 rank score for

the neighborhood of birth. We are careful to match only on characteristics that occur early in life (by 18 months

of age) and that could not result from lead exposure. CEM is able to match 85% of students that have EBLLS in

the recent cohort and 82% in the early cohort with students in their own cohort with BLLs that are underneath the threshold.

Using these matched groups produced through CEM, we then estimate a weighted linear regression

model for each outcome of interest with the dichotomous measure of having EBLL or not as the sole regressor.

The coefficients derived from these models are used to evaluate whether there is a statistically significant im-

pact of lead poisoning on the outcomes and to estimate the percentage differences in the rates between the

groups that can be attributed to lead.

FINDINGS

The study shows that there is a large impact of lead poisoning on children, and that the disparities be-

tween the students with and without elevated lead levels first seen in early childhood persist through early

adulthood. The results of the regression models confirm that the children with EBLL compared to the matched

comparison children without EBLL have significantly worse outcomes across multiple systems. Table 2 displays

the estimated parameters derived from OLS regression models applied to coarsened-exact-matched groups,

where 'outcome’ is the dependent variable and EBLL is the binary independent variable. These parameters rep-

resent the estimated percentage of children having each outcome in the group without EBLL (base rate) and the

percentage point difference from this base rate in the EBLL group. We also provide the effect size due to lead

expressed as a percent difference (% Δ). For example, in the first row of the table we see that the rate of being

on track according to the KRA-L test is estimated at 33 percent for the group without EBLL in the recent cohort.

For the group with EBLL, the KRA-L on-track rate is 9% lower. The estimated impact of lead (% Δ) is -27%, which

means that children with EBLL have a 27% lower chance of being on-track for kindergarten than children with-

out EBLL.


6

*Coefficients derived from OLS regression models applied to coarsened-exact-matched groups, where 'outcome’ is

the dependent variable and EBLL is the binary independent variable. ** This is the estimate of model constant (no

EBLL). ***This is the estimate of the coefficient on the dependent variable (with EBLL). NS denotes impact is not sta-

tistically significant.

Table 2. Impact of EBLL on selected outcomes (Model based estimates*)

Outcomes

2007-2008 9th grade students 2016-2017 9th grade students

Estimated percent in

group with-out EBLL **

Estimated percent

point differ-ence in EBLL

group***

Impact of lead- effect size ( % Δ )

Estimated percent in

group with-out EBLL

Estimated percent

point differ-ence in EBLL

group

Impact of lead- effect size

( % Δ )

Education

KRA_L (on-track) 32.9 -8.8 -27%

3rd GR reading (Pass) 34.8 -11.0 -32%

3rd GR math (Pass) 55.5 -6.2 -11%

6th GR reading (Pass) 60.7 -9.6 -16%

6th GR math (Pass) 42.1 -5.9 -14%

8th GR reading (Pass) 59.4 -11.4 -19% 31.6 -6.8 -21%

8th GR math (Pass) 39.1 -7.5 -19% 25.6 -6.5 -25%

OGT Reading (Pass) 73.3 -11.7 -16% 36.1 -8.3 -23%

OGT Math (Pass) 62.7 22.2 -4.7Grade repetition in 3rd 1.7 2.7 0.6 (NS)

Grade repetition at 9th 23.2

-12.10.7(NS)7.1

-19%

(NS)31% 12.8 3.5

-21%

(NS)27%

High school grad on-time 51.8 -9.2 -18%

High school grad (ever) 59.1 -7.5 -13%

College matriculation 55.2 -9.8 -18%

Judicial system

Delinquency filing (any) 21.7 3.5 16% 14.8 4.0 27%Delinquency filing (violence) 11.6 3.2 27% 9.0 3.7 41%

Adult incarceration (age 18-23) 18.0 6.4 35%

Homeless Services at age 18-23

Homeless service (any) 3.8 1.7 45%

Emergency shelter (any) 1.9 1.4 73%

Public assistance at age 23

TANF (1 month or more) 3.8 1.8 49%

SNAP (2qts or more) 28.6 4.9 18%

Employment at age 23

Employment (2qts or more) 62.0 -2.7 (NS) (NS)


7

Moving onto grades 3-6, Table 2 also shows large differences between the group without EBLL and the

EBLL group on proficiency test results. For example, in the recent cohort, the rates of passing the third grade

reading and math proficiency tests are 32% and 11% lower respectively for the EBLL group compared to the

group without EBLL. The proficiency test passing rates continue to be significantly depressed for the EBLL group

relative to the group without EBLL through 6th grade, although at somewhat lower magnitudes than in earlier

grades.

From the 8th grade forward, we have testing data for both the early and recent cohorts. Reading and

math proficiency passage rates are significantly lower for the EBLL children compared to the group without

EBLL. The impact estimates range from 19% to 25% lower depending on the test. The Ohio Graduation Test

passage rates are significantly lower for the EBLL group in both cohorts, and the percentage difference is of

similar magnitude. However, the overall passage rate on OGT is higher for the early cohort than the recent

cohort, re-flecting changes the state made to the OGT test 2009.

Table 2 also displays grade repetition rates for both cohorts. Grade repetition is a low frequency event in

general, but is costly to the school system and to the child. We see significant differences in repetition rates be-

tween the groups in the third grade (early cohort only) and ninth grade, with EBLL children repeating these

grades more often. Repetition rates for other grades are very low, and the differences between groups are not

statistically significant.

Lead poisoning also has an impact on outcomes in the juvenile justice systems as shown in Table 2. In

both cohorts, youth in the EBLL group well exceed the group without EBLL on all juvenile court filings and on

those specifically for violent incidents. Although the incidence rates of overall juvenile delinquency for the non-

exposed groups decline in the recent cohort relative to the early cohort, the disparities between EBLL and non-

EBLL groups (Δ) remain similar across cohorts. However, disparities between the groups on violent crimes have

increased in the recent cohort translating into a 27% and 41% effect size (%Δ) respectively. Table 2 also displays

some early adult outcomes at ages 18-23 for the early cohort. The rates of college matriculation in Ohio colleges

are lower in the EBLL group compared to the group without EBLL. Moreover, the group with EBLL has much

higher incarceration rates and likelihood of using homeless services and homeless shelters than the group with-

out EBLL. At age 23, individuals in the EBLL group are more likely to have relied on public assistance programs

such as TANF (for at least one month) and SNAP (for at least half the year). There was no statistically significant

impact of lead poisoning on working in covered employment for at least two quarters at age 23.

CONCLUSION

This study examined the impact of early childhood lead poisoning on outcomes throughout childhood

and early adulthood. By comparing carefully matched groups of children with and without elevated blood lead

levels, this research demonstrates the large impact of lead poisoning on individuals and public systems. Across

many outcomes, the disparities between individuals who had elevated blood lead levels in early childhood, com-

pared to their matched counterparts without EBLLs, are in the ranges of 20-40%. All of the outcomes examined

in this study are costly to the individuals who experience them, the systems that serve them and society at-

large.

Our analysis also shows that black or African American youth are disproportionately poisoned by lead as

young children and that this leads to a series of other disadvantages as they grow up. It is worth noting that

some of the largest impacts of lead poisoning are in justice system involvement, where other factors such as


8

societal biases, policing and judicial practices play a role in who gets into these systems. In this study, the EBLL

and non-EBLL groups are matched on race and numerous other characteristics before age 3 in order to isolate

the effects of lead poisoning from other confounding factors. However, it is important that future work explore

within-race effects of lead poisoning with the aim of further disentangling the intrinsic impact of lead from ex-

ternal and systemic factors related to systemic racism.

In interpreting the results of this longitudinal study, it is useful to consider that early development and

experiences set the stage for individuals’ subsequent progress. For example, being less ready for kindergarten

puts children at greater risk for not being proficient in reading and math during elementary school, needing to

repeat a grade, and later difficulties in educational attainment and labor market success. Similarly, behavioral

problems in early childhood can persist and contribute to later antisocial behavior, social dislocations and barri-

ers to employment. From this longitudinal perspective, lead poisoning in early childhood can shift the trajectory

at various developmental stages and have long-term consequences for the individual.

In considering societal costs, we see a similar cumulative effect as children with EBLLs move through

various systems. When they are in elementary school, lead poisoned children have lower passage rates on profi-

ciency exams and higher rates of grade retention, necessitating higher levels of educational supports and ser-

vices. These needs persist in high school, as evidenced by increased rates of grade repetition and more difficulty

in passing graduation tests. From the perspective of the justice system, the impact can be seen in the increased

rates of involvement in both the juvenile justice and adult corrections systems that are attributable to elevated

lead levels in early childhood. Other social service systems are also affected by lead poisoning, through in-

creased utilization of homelessness services and public assistance as adults among individuals exposed to lead

as children. Although beyond the scope of this study, the educational deficits and justice system involvement

induced by lead poisoning are likely to play out over adulthood in the form of lost human capital and lifetime

earnings. This study underscores the urgency of implementing lead poisoning prevention programs that reduce

societal costs and move us towards a more equitable and just society.

REFERENCES

1. Gould, E. (2009). Childhood lead poisoning: Conservative estimates of the social and economic benefits of

lead hazard control. Environmental Health Perspectives. 117, 1162–1167.

2. Landrigan, P. J., Schechter, C. B., Lipton, J. M., Fahs, M. C. & Schwartz, J. (2002). Environmental pollutants

and disease in American children: Estimates of morbidity, mortality, and costs for lead poisoning, asthma,

cancer, and developmental disabilities. Environ. Health Perspect. 110, 721–728 (2002).

3. Pocock, S. J., Smith, M. & Baghurst, P. (1994). Environmental lead and children’s intelligence: A system-

atic review of the epidemiological evidence. BMJ 309, 1189.

4. Lanphear, B. P., Dietrich, K., Auinger, P. & Cox, C. (2000). Cognitive deficits associated with blood lead

concentrations < 10 μg/dL in US children and adolescents. Public Health Rep. 115, 521–529.

5. Canfield, R. L. Henderson, C. R. Cory-Slechta, D. A. Cox, C., J., Jusko, T.A., & Lanphear, B. P. (2003). Intel-

lectual impairment in children with blood lead concentrations below 10 μg per deciliter. N. Engl. J. Med.

348, 1517–1526.

6. Amato, M. S., Moore, C. F., Magzamen, S., Imm, P., Havlena, J. A., Anderson, H. A., & Kanarek, M. S.

(2012). Lead exposure and educational proficiency: moderate lead exposure and educational proficiency

on end-of-grade examinations. Ann. Epidemiol. 22, 738–743.


9

7. Zhang, Na., Baker, H. W., Tufts, M., Raymond, R. E., Salihu, H., & Elliott, M. R. (2013).Early childhood lead exposure and academic achievement: Evidence from Detroit public schools, 2008-2010. Am. J. Public Health 103.

8. Marcus, D. K., Fulton, J. J. & Clarke, E. J. (2010). Lead and conduct problems: A meta-analysis. J. Clin. Child Adolesc. Psychol. 39, 234–241.

9. Wasserman, G. A., Staghezza-Jaramillo, B., Shrout, P., Popovac, D. & Graziano, J. (1998). The effect of lead

exposure on behavior problems in preschool children. Am. J. Public Health 88, 481–486.

10. Needleman, H. L. & Gatsonis, C. A. (1990). Low-Level Lead Exposure and the IQ of Children A Meta-

analysis of Modern Studies. JAMA 263, 673–678.

11. Coulton, C., Richter, F., Kim, S., Fischer, R. & Cho, Y. (2016). Temporal effects of distressed housing on early childhood risk factors and kindergarten readiness. Child. Youth Serv. Rev. 68, 59–72.

12. Anthony, E., Steh, S., Atwell, M. S. & Fischer, R. (2019). Early Childhood Lead Exposure in Cuyahoga Coun-

ty and the Impact on Kindergarten Readiness. Cleveland, OH: Mandel School of Applied Social Sciences, Case Western Reserve University. January.

13. Reuben, A., Caspi, A., Belsky, D.W., Broadbent, J., et. al. (2017). Association of Childhood Blood Lead Lev-

els With Cognitive Function and Socioeconomic Status at Age 38 Years and With IQ Change and Socioeco-

nomic Mobility Between Childhood and Adulthood Supplemental content. JAMA 317, 1244–1251.

14. Shadbegian, R., Guignet, D., Klemick, H. & Bui, L. (2019) Early childhood lead exposure and the persis-

tence of educational consequences into adolescence. Environ. Res. 178, 108643

15. Grönqvist, H., Nilsson, J. P. & Robling, P.-O (2017). Early lead exposure and outcomes in adulthood. A Ser-

vice of zbw Leibniz-Informationszentrum Wirtschaft Leibniz Information Centre for Economics Standard-

Nutzungsbedingungen.

16. Aizer, A. & Currie, J. (2019). Lead and Juvenile Delinquency: New Evidence from Linked Birth, School, and Juvenile Detention Records. Rev. Econ. Stat. 101, 575–587.

17. Sampson, R. J. & Winter, A. S. (2018). Poisoned development: assessing childhood lead exposure as a cause of crime in a birth cohort followed through adolescence. Criminology 56, 269–301.

18. Stefano M., King, G. & Porro, G. (2012). Causal inference without balance checking: coarsened exact matching. Political Analysis 20:1–24.

19. Chetty,R, Friedman,J.N. Hendren, N., Jones, M. R., & Porter, S. R. (2018) The Opportunity Atlas: Mapping the Childhood Roots of Social Mobility. NBER Working Paper 25147: 18-42


10

APPENDIX A:

Note. * Used capillary and venous measure (E) Early cohort only; (R) Recent cohort only; (ER) Early and Recent cohorts

Concept Measure Source

Predictor Confirmed EBLL ODH-L

Downstream outcomes Achievement test SD+EMIS

Grade repetition

Geometric mean of lead tests until 60 months is at or above 5 BLL (μg/dl)(Yes= 1)

Kindergarten Readiness Assessment-Literacy score is on track (R) Proficiency tests of Math, Reading at 8th grade(E), at 3, 6, 8th grade(R)

Ohio Graduation Test of Math, Reading in high school (ER)

Grade repetition at 3-9th grade (ER) (Yes=1) SD+EMIS

High school graduation High school graduation on-time and ever (E) (Yes=1) SD+EMIS

College matriculation College matriculation at age 18-23 (E) (Yes=1) SD+EMIS

Juvenile delinquency Delinquency court filing (All or only violence) ever at age 9-16(ER) (Yes=1) CC JC

Adult Incarceration Adult incarceration ever at age 18-23(E) (Yes = 1) CC SO

Homelessness Homeless services or shelter use ever at age 18-23(E), at age 14-16(R) (Yes=1) CC HMIS

Public assistance Receiving TANF ever; SNAP 2 quarters or more at age 18-23 (E) (Yes=1) CC JFS

Employment Employed in at least 2 quarters at age 23 (Yes=1) (E) ODJFS

Child characteristics DOB Date of Birth (MM/ DD/ YYYY) ODH-B

9th grade entry School year entering CMSD 9th grade (ER) SD

Gender Female=1/ Male=0 ODH-B

Race/ethnicity African American/ White/Hispanic/Other (Yes=1) ODH-B

Low birth weight Low birth weight (<2,500 grams; Yes=1) ODH-B

Premature birth Premature (<37 weeks gestation; Yes=1) ODH-B

Apgar score 5 minute Apgar score (0-10) ODH-B

Child maltreatment Child neglect/abuse investigation (Yes=1) CC DCFS

Foster care Foster care placement (Yes=1) CC DCFS

Family characteristics Teen mother Born to a teen mother (Yes=1) ODH-B

Mother’s education Born to a mother with high school diploma (Yes=1) ODH-B

Marital status Non-married=0/ Married=1 at child birth ODH-B

Prenatal care Kessner's Index (Adequate; Yes=1) ODH-B

Risky health behavior Tobacco use during pregnancy (Yes=1) ODH-B

Alcohol use during pregnancy (Yes=1) ODH-B

Native language Student's native language is NOT English (Yes=1) EMIS

Housing assistance Ever Received Housing voucher or public housing (12-18 months) in early child-hood (Yes=1)

CMHA

Public assistance Ever Received TANF, SNAP, Medicaid (12-18 months) in early childhood (Yes=1) CC JFS

Neighborhood characteristics Neighborhood Census tract at child birth ODH-B+C

Opportunity index** Opportunity Atlas, census tract level, 2014-2015 ODH-B+C+T


11

**The opportunity index ranks neighborhoods by the average income level of individuals between the ages of 31 and 37

who grew up in that neighborhood and in low income families. For a detailed explanation of the index see Raj Chetty, John

N. Friedman, Nathaniel Hendren, Maggie R. Jones & Sonya R. Porter, 2018. "The Opportunity Atlas: Mapping the Childhood

Roots of Social Mobility," NBER Working Paper 25147: 18-42, Center for Economic Studies, U.S. Census Bureau.

Sources. CC JC: Cuyahoga County Juvenile Court CC SO: Cuyahoga County Sheriff’s Office CC DCFS: Cuyahoga County Division of Children and Family Services CC JFS: Cuyahoga County Job and Family Services CC HMIS: Cuyahoga County Homeless Management Information System CMSD: Cleveland Metropolitan School District SUB: Inner ring suburban School District SD: CMSD + SUB CMHA: Cuyahoga Metropolitan Housing Authority ODH: Ohio Department of Health (B=birth; L=lead) EMIS: Ohio Educational Management Information System ODJFS: Ohio Department of Job and Family Services C: 1990, 2000, 2010 Decennial Census, 2005-2015 American Community Survey (ACS) T: Federal income tax returns for 1989, 1994, 1995, and 1998-2015 (opportunityatlas.org)

The Center on Urban Poverty and Community Development at Case Western Reserve University is committed to high standards of data and research integrity. We carefully follow human subject research standards, which includes Institutional Review Board (IRB) oversight, and we are dedicated to providing criticaldata analysis as well as nuanced presentation to protect against inaccuracies. However, even adherence tohigh standards of data integrity has limitations. All administrative data has the potential for bias due to theproven history of over-surveillance and unequal treatment of poor communities, particularly communitiesof color. We acknowledge that these biases could impact the administrative data that we receive. Our Centeris guided by these considerations as we work to provide an ethical and accurate reflection of the communities withwhich we work.

downstream consequences of childhood lead poisoning€¦ · lead poisoning in young children has...

Documents